Many survivors of brain injury have persistent impairment of hand function despite receiving conventional therapy. The long-term goal of this research project is to augment and direct the brain's inherent plasticity to improve motor function for survivors of traumatic or ischemic brain injury. Functional improvement after brain injury correlates with an enlarged area of cerebral cortex corresponding to the improved movement (motor map), but it is unclear if the enlarged map causes improved function. This question is of fundamental importance to our understanding of recovery from brain injuries. Brain machine interfaces (BMIs), which enable subjects to use their brain signals to directly control external devices, can induce plastic changes in the brain's activity. Thus, a BMI could provide a powerful tool to drive plasticity in injured brains and also test the effects of map enlargement on function. However, important gaps in our knowledge remain about what aspects of BMI training are critical to enhancing cortical plasticity, including 1) the types and features of neural signals used to control the BMI, 2) the temporal precision with which somatosensory (haptic) feedback must be synchronized with motor intent, and 3) the spatial precision of movement intent (e.g., individual finger vs. whole hand) used to control the BMI. The objectives of this proposal are to determine the aspects of BMIs most important to changing motor maps, and the extent to which motor map expansion driven by BMIs improves function. By expanding the map, the control of muscles that have by paralyzed by brain injury can be moved to areas of cortex that still retain intact descending connections, thus restoring function. The central hypothesis of this proposal is that optimally driving plasticity and motor map changes is critically dependent on simultaneously activating motor intent and haptic feedback. We propose that high-frequency signals will enable much greater spatiotemporal precision than the low frequencies used in BMIs for rehabilitation to date. We will test this hypothesis in subjects who have had hemicraniectomies for traumatic brain injury via these specific aims: 1) Determine the extent to which high-frequency based BMI training drives motor map enlargement and improves hand function, and 2) Determine the role of synchrony between motor intent and haptic feedback in driving changes in map size and hand function. This proposal's innovative use of scalp signals over the hemicraniectomy will enable us to record high-resolution, high-bandwidth signals non-invasively and test our hypothesis. Achieving our objectives will be significant because it will improve the design of BMI training paradigms by identifying spectral, temporal, and spatial features that are critical to plasticity enhancement. We expect this proposal to define the ability of BMIs to create changes in motor maps. We also expect it to help define the relationship between motor map changes and functional recovery. Finally, it will demonstrate the potential for functional improvement in future studies using minimally-invasive, epidural-based BMIs after brain injury.